Bots trained to play like a human are more fun
Authors:
- Bhuman SoniBSBhuman Soni
- Philip HingstonPHPhilip Hingston
- 1 June 2008
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 21614393,p. 363-369
Abstract
Computational intelligence methods are well-suited for use in computer controlled opponents for video games. In many other applications of these methods, the aim is to simulate near-optimal intelligent behaviour. But in video games, the aim is to provide interesting opponents for human players, not optimal ones. In this study, we trained neural network-based computer controlled opponents to play like a human in a popular first-person shooter. We then had gamers play-test these opponents as well as a hand-coded opponent, and surveyed them to find out which opponents they enjoyed more. Our results show that the neural network-based opponents were clearly preferred.Keywords
This publication has 4 references indexed in Scilit:
- Exploiting Intelligence in Fighting Action Games Using Neural NetworksIEICE Transactions on Information and Systems, 2006
- Capturing Entertainment Through Heart Rate Dynamics in the Playware PlaygroundLecture Notes in Computer Science, 2006
- Machine learning techniques for FPS in Q32004
- Learning Human-Like Opponent Behavior for Interactive Computer GamesLecture Notes in Computer Science, 2003